GNOME Bugzilla – Bug 449272
Screensaver blanks screen without ability to restore
Last modified: 2009-07-27 23:21:59 UTC
[ forwarded from http://bugs.debian.org/429422 from Brian Sipos ] When the screensaver is fading out, a mouse movement will cause the screensaver to be disabled. Once the fade-out has occurred, a mouse movement will promopt for a password to restore display. If a mouse movement occurs at some time when the screen has completely darkened from the fading for some reason the password prompt will never display and it is impossible to restore the display without killing the X server and logging-in again. This is a very severe problem and I have experienced this issue several times now. I can attempt to repeat the issue, but so far it has occurred at inopportune times and is very annoying. 8< --- 8< --- This kind of bug (with some gamma issues making the screen entirely black) occured very frequently in early releases, and it went away some time ago. However there are still some cases where it happens; it has happened to me at least once with the 2.18 series.
Seems x86_64 specific; I never got the bug and the two bug submitters run x86_64.
One submitter noticed that /usr/lib/gnome-screensaver/gnome-screensaver-dialog segfaults for him and sent us two backtraces: :~$ gdb /usr/lib/gnome-screensaver/gnome-screensaver-dialog [...] [New Thread 47425885751824 (LWP 8115)] (gnome-screensaver-dialog:8115): Gnome-CRITICAL **: gnome_program_locate_file: assertion `program != NULL' failed Program received signal SIGSEGV, Segmentation fault.
+ Trace 150903
Thread 47425885751824 (LWP 8115)
Reassigning to libgnomekbd as the crash doesn't seem to be related to gnome-screensaver in particular.
I think that this a encoding-related issue, see bug #529773 to see another backtrace and my explanation of the problem.
In spite of what I said in my previous comment, after reading more carefully the description of this bug I think that it may be unrelated to bug #529773. However as that one has just been fixed it would be a good idea if the reporter could try again with a patched version of libxklavier to confirm this. Thanks!
Well, since it is about a year old, I a closing it.